hchang 23 Nov 2021
Face verification 얼굴 인식, 구분
inspired by curriculum learning
difficulty of triplets
Rekated Work 얼굴 인식 분야에서 현 상황의 리뷰?
Method
3.1
The embedding is represented by f(x) ∈ R d . It embeds an image x into a d-dimensional Euclidean space. Additionally, we constrain this embedding to live on the d-dimensional hypersphere, i.e. kf(x)k2 = 1. This loss is
$L =XNihkf(xai) − f(x^p_i)k22 − kf(x^a_i) − f(x^n_i)k22 + αi_+$
보면 수식 설명이 쭉 나옵니다. 여기서 중요한 부분은 아래 내용입니다. (물론 수식도 중요하지만요.)
Generating all possible triplets would result in many
triplets that are easily satisfied (i.e. fulfill the constraint
in Eq. (1)). These triplets would not contribute to the training
and result in slower convergence, as they would still
be passed through the network. It is crucial to select hard
triplets, that are active and can therefore contribute to improving
the model. The following section talks about the
different approaches we use for the triplet selection.
바로 적절한 triplet을 선택하는 일입니다.
This
means that, given $x^a_i$,
we want to select an $x^p_i$(hard positive)such that
$argmax_{x^p_i} ||f(x^a_i) − f(x^p_i)||^22$
and similarly
$x^n_i$(hard negative) such that
$argmin{x^n_i} ||f(x^a_i) − f(x^n_i)||^2_2$
얼굴의 이미지를 유사도를 나타내는 compact한 유클리디안 공간으로 옮겨준다.